Aggregating Versus Blending: What’s the Difference?

Intellicast
Intellicast S4E4 – Xcelerant, Behaviorally, and New Legislation
January 25, 2021
Intellicast S4E5 – Stock Market Craziness, Qualtrics IPO, and More!
February 2, 2021
Show all

Aggregating Versus Blending: What’s the Difference?

Sample blending. You’ve probably heard the term, but what does it mean?

Before we answer that, we need to give you a bit of history. Years ago, it was a best practice in online quantitative research to use a single sample panel for your study. The belief was that to maintain consistency, everything had to remain the same, including the sample.

Over the last few years though, this belief has been proven wrong.

Whether by necessity or evolved thinking, the use of a single sample source has declined. Now the best practice for studies, including wave and tracking studies, is to blend sample from multiple sources.

This may sound great, but did you know there are not only different ways to blend sample, but some are better than others? Probably not.

The first way is to aggregate sample providers. With this method, you are really adding additional panels because one source cannot yield enough completes. So, you add another, and another, until you reach the desired base size. You will find this method used a lot when a study needs to be topped-up. Think stacking rocks on one another – that’s what you are doing with sample providers. This method assumes all panels are the same and are interchangeable. (Hint – they’re not.)

The second method is to blend sample. With this method, you plan to use 3 or more sample providers from the start, selecting them so that you not only have plenty of feasibility, but also a reduced field time and a reduced risk of sample bias. This method allows you to consider that not all panels are the same. They can provide different feedback because of the following:

  • Recruiting methods
  • Incentives
  • Industry consolidation
  • Attrition/Turnover
  • Investment changes
  • Ways to access the survey (portal, email, pop-ups)
  • Attitudinal and behavioral differences
  • Client needs (panels aren’t built to match census)

There is also a third way – strategic sample blending. This method builds on the blending sample approach but takes into account the behavioral and attitudinal differences of panels and blends them in an intentional and controlled method. This allows the insights gathered to be more representative and accurate. If you are looking for a best practice in blending sample, you can learn more about IntelliBlend, EMI’s patented strategic sample blending methodology leveraging our proprietary research-on-research.